具備波束成型與功率控制技術之第五代行動通訊與WiFi網路共存的設計與實現

Abstract

近年來隨著科技不斷進步，人們對於連網需求日益增加，但是同一時間內頻譜資源是有限的，因此電信商需要取得更多頻譜資源以服務使用者，加上2020年美國聯邦通訊委員會 (Federal Communication Commission, FCC) 公布了位於6 GHz，頻寬為1200 MHz的未授權頻帶供給Wi-Fi技術以及其他使用者使用，所以將以往只使用在授權頻帶上的通訊技術拓展到未授權頻帶上使用漸漸受到重視，在未授權頻帶上通訊雖然能夠獲取更多頻譜資源，但須要面臨受其他通訊技術干擾的風險，例如藍芽、Wi-Fi等，而FCC公布的頻帶主要給Wi-Fi技術使用，因此如何與Wi-Fi技術共存儼然成為重要議題，本論文引入空間多工 (Spatial Multiplexing) 概念，透過波束成型技術降低5G NR對Wi-Fi技術的干擾，讓兩系統在碰撞情況下也能夠各自解調成功，達成5G NR、Wi-Fi在同頻帶共存。 第二章首先分別介紹5G NR以及Wi-Fi 6標準，接著簡介何謂未授權頻譜以及現有使用在未授權頻譜上的通訊技術，最後就是介紹本論文主要使用的最小變異量無失真響應 (Minimum Variance Distortionless Response, MVDR) 波束成型演算法。 第三章我們說明基於5G新無線電 (New Radio, NR) 標準所實作的八天線具波束成型硬體發射機，並將其整合在賽靈思射頻系統單晶片，搭配MVDR波束成型演算法，在集中訊號給5G UE的同時降低干擾Wi-Fi終端 (Station)。5G接收機則是使用國家儀器軟體無線電 (NI USRP 2943R) 搭配四天線陣列以及合成 (Combining) 技術，我們透過真實空氣通道 (Over The Air, OTA) 驗證其相比單天線可以提高4 ~ 5 dB SINR。最後為了驗證MVDR波束成型在共存系統之可行性，我們成功在OTA環境下達成5G、Wi-Fi同一時間在同頻段可以各自服務兩個使用者，總體峰值傳輸率 (Peak Data Rate) 達到162.33 Mbps。 第四章中，我們將說明本文透過窗函數進行改良的MVDR演算法，並且實際在真實通道下比較不同窗函數的空間濾波特性，包括主波瓣強度，旁瓣抑制程度等，最後是將其實際應用在5G、Wi-Fi共存系統中，並在不同情境下比較各種窗函數MVDR波束成型係數之頻譜效率。 第五章我們將5G NR-U、Wi-Fi系統共存更貼近真實情境，我們在5G基地台假設Wi-Fi接收機角度未知，因此設計波束成型係數前需要透過STA上行訊號估計到達方位角 (Angle of Arrival Estimation, AoA Estimation)，我們使用一片Xilinx RFSoC開發四天線接收機，透過RFSoC接收STA上行訊號後，傳回Host PC估計方位角，再根據估計之方位角計算波束成型係數。除此之外，我們在5G NR-U基地台引入功率控制技術， 動態調整5G發射功率以維持系統間共存，同時我們針對傳統的固定步階功率控制 (Fixed-step Power Control) 演算法穩定性不足的問題進行修正，提出了基於歷史的 (History-Based) 固定步階功率控制，相比在5G基地台只使用波束成型技術，最多可以提升5G能源效率75%、Wi-Fi能源效率83%，最終我們在空氣通道下驗證了具備共存裝置追蹤與功率控制技術之5G NR-U基地台，除了能夠在1.2秒內追蹤共存裝置外，透過功率控制技術能夠在Wi-Fi AP發射功率改變下維持兩系統共存。

In recent years, as technology advances, people’s demand for connectivity is increasing. However, spectrum resources are limited at the same time, so telecom operators need to obtain more spectrum resources to serve users. In 2020, FCC announced the unlicensed band at 6 GHz with a bandwidth of 1200 MHz for Wi-Fi technology and other users. As a result, there is a growing interest in expanding the use of communication technologies that were previously only in licensed frequency bands to unlicensed bands. Communication in an unlicensed band provides access to more spectrum resources, but runs the risk of interference from other communication technologies, for example, Wi-Fi, Bluetooth, etc. Because Wi-Fi primarily uses the FCC-announced unlicensed band, coexistence with Wi-Fi technology has become an important topic. This paper introduces the concept of Spatial Multiplexing to mitigate the impact of 5G NR on Wi-Fi via beamforming technology, allowing both systems to successfully demodulate under collision and achieve the coexistence of 5G NR and Wi-Fi in the same frequency band. Chapter 2 will first introduce the 5G and Wi-Fi 6 standards. We introduce the unlicensed spectrum and the existing technologies used in the unlicensed spectrum, and finally, we present Minimum Variance Distortionless Response (MVDR) beamforming algorithm used in this paper. In Chapter 3, we explain the implementation of an eight-antenna beamforming hardware transmitter based on the 5G New Radio standard. It was integrated into Xilinx RFSoC with the MVDR beamforming to reduce interference with Wi-Fi receivers while concentrating the signals to the 5G UE. The 5G receiver uses the National Instruments Software Defined Radio (NI USRP 2943R) with a four-antenna array and combining technique. It is verified in the Over the Air (OTA) to improve SINR by 4 to 6 dB over a single antenna. Finally, to validate the feasibility of MVDR beamforming in the coexistence system, we successfully achieved 5G and Wi-Fi in the OTA, each of which can serve two users in the same frequency band simultaneously, with an overall peak data rate of 162.33 Mbps. In Chapter 4, we will introduce the improved MVDR algorithm by window functions, and compare the spatial filtering properties of different window functions in the OTA, including the main lobe power and sidelobe suppression, etc. Finally, we will apply it to the 5G and Wi-Fi coexistence system and compare the spectrum efficiency of various window function MVDR beamforming weights under different scenarios. In Chapter 5, we bring the coexistence of 5G NR-U and Wi-Fi systems closer to the actual situation. We assume the Wi-Fi STA’s azimuth angle is unknown at 5G BS. As a result, before calculating beamforming weights, we must estimate the Angle of Arrival of the Wi-Fi STA using the STA uplink signal. Therefore, we develop a four-antenna receiver with one Xilinx RFSoC. A four-antenna receiver receives the uplink signal sent by Wi-Fi STA and estimates the Angle of Arrival by the host PC. Then, the host PC recalculates beamforming weights based on the estimated azimuth angle. Furthermore, we introduce power control techniques in the 5G NR-U BS to dynamically adjust 5G Tx power to maintain system coexistence. To address the lack of stability of the traditional fixed-step power control algorithm, we proposed history-based fixed-step power control. Compared to using only beamforming technology in the 5G NR-U base station, the 5G energy efficiency is up by 75%, and the Wi-Fi energy efficiency is up by 83%. Ultimately, we validated the 5G NR-U BS with Wi-Fi device tracking and power control technology in OTA. In addition to tracking the coexistence device in 1.2 seconds, the power control technology can maintain the coexistence between the two systems as the Wi-Fi AP transmitting power changes.